Sunday, April 29, 2007

Physics intuition and a follow-up

I just had two interesting experiences. First, I spent a couple of hours reading a PhD thesis on a topic that had a strong tie to nonlinear dynamics and chaos. It had all the fun stuff: Poincare sections, phase space localization, etc., in the context of classical elliptical orbits with "kicks" applied as drive. While reading this, I had the realization that I had very little physical intuition for this system, even though it's in many ways an old problem. For example, the statement that, in this 1/r^2 central force problem, trajectories with large angular momentum have less orbital eccentricity did not seem obvious to me - I really had to think about it. Why do I have more intuition for nanoscale and quantum systems than classical central force problems? Because I hardly ever work on the latter. Physical intuition is the intellectual equivalent of muscle mass in some specific group. If I don't exercise the Poisson bracket/Runge-Lenz vector part of my physics brain, it atrophies.

The second experience also relates to intuition. Remember this post? PRL followed up with me last August. They said that they'd looked into my concerns about data manipulation, and that the author (not clear which one they contacted) had shown them "unprocessed" data, and that things looked ok to them. Well, I've been contacted by a colleague at another institution who read my blog post about this, figured out which paper I meant (!), and alerted me to other questionable figures in other publications. Updates as events warrant.

Thursday, April 26, 2007

This week in cond-mat

Briefly emerging from my end-of-semester fog, here are some interesting preprints from the past week.

The graphene craze continues unabated. Remember how the superconductivity community descended upon MgB2 and made every superconductivity-related measurement under the sun on the new material in a feeding frenzy? A similar phenomena is taking place with the 2d electron community and graphene. Fortunately, graphene seems to be pretty neat stuff! For example:
arxiv:0704.3165 - Hill et al., Graphene spin valve devices
People have done normal metal contacts to graphene, and superconducting contacts to graphene, so what's left but ferromagnetic contacts to graphene? Unsurprisingly you can use ferromagnetic electrodes to inject spin into graphene, and its such a low-Z material of high purity that both spin-orbit scattering and spin flip scattering from impurities are minimal, leading to real spintronic possibilities in this stuff.

Further exploiting the robustness of graphene even under significant processing:
arxiv:0704.2626 - Huard et al., Transport measurements across a tunable potential barrier in graphene
arxiv:0704.3487 - Williams et al., Quantum Hall Effect in a graphene pn junction
arxiv:0704.3608 - Abanin and Levitov, Quantized transport in graphene pn junctions in magnetic field
Because graphene is a high quality 2d material and can be shifted readily from n and p carriers via doping or gating, it is possible to set up sophisticated structures (npn or pnp junctions; pn junctions) while preserving long mean free paths. The result is rich phenomenology, as seen in the first two (experimental) papers listed here, and analyzed in detail in the third (theory) paper. I'm still waiting for a really unexpected graphene result that isn't readily explained.

Two other papers that involve tunable model systems to examine strong correlation physics:
arxiv:0704.3011 - Bloch et al., Many-body physics with ultracold gases
This is a review article about using cold atoms to look at nontrivial correlation effects. One holy grail in this business is to use strongly interacting cold fermions in a 2d optical lattice to explicitly simulate the Hubbard model (relevant to high-Tc superconductivity), a topic of much interest to one of my faculty colleagues.

arxiv:0704.2614 - Walsh et al., Screening of excitons in single, suspended carbon nanotubes
Carbon nanotubes have 1d band structures, and therefore are subject to strong electron-electron interaction effects and poor screening. The consequence of these interactions is the demise of Fermi liquid theory, and therefore the onset of the fractionalized quasiparticles (spinons and holons) of Luttinger liquid theory. Excitons are also strongly modified in these systems. One way to probe these effects is to change the effective interaction; this is done by using immersion in dielectric media to change the screening of charges, and the effects are probed spectroscopically.

Monday, April 16, 2007

This week in cond-mat, self-indulgent edition

Two cond-mat papers this week, and one other item.

arxiv:0704.1775 - Yu et al., Origin of discrepancies in inelastic electron tunneling spectra of molecular junctions
This paper is by my former student, Lam Yu, now at NIST in Maryland. It is a systematic study of a relatively long-standing problem in the molecular electronics field. Electrons can undergo off-resonant tunneling through molecules; that is, the electrons tunnel from one electrode to the other via the molecule, even though there are no molecular levels aligned with the filled electronic states of the electrodes. You can think of this as a second-order tunneling process: while actually putting an extra electron on the molecule is classically forbidden by conservation of energy, one can consider a second order tunneling process where the electron-on-the-molecule is a virtual intermediate state. If a sufficient bias voltage exists between the source and drain electrodes, and the nuclear wave functions work out right (i.e., Franck-Condon factors), one can have processes where the electron tunnels on to the vibrational ground state of the molecule and tunnels off a vibrationally excited state of the molecule, all in one process. This is the key to inelastic electron tunneling spectroscopy, where such vibrational excitations result in a signature in the IV characteristics of the electrode/molecule/electrode sandwich (nominally a peak in d^2I/dV^2). The problem is, experiments have shown a variety of lineshapes rather than simple peaks. Lam's work shows that the presence of metal ions within the molecular layer can result in complicated lineshapes like those seen in some experiments.

arxiv:0704.0451 - Ward et al., Electromigrated nanoscale gaps for surface-enhanced Raman spectroscopy
This paper comes out of my own group. It has been known for some time that metal nanostructures driven at their plasmon resonances can act like little optical antennas, so that the local electromagnetic field near, e.g., metal nanoparticles can be much larger than the incident electromagnetic field. Since Raman scattering (a common vibrational spectroscopy) is a nonlinear optical process that scales like (roughly) the fourth power of the electric field, it is a prime candidate for these enhancement effects. In some geometries, single-molecule Raman sensitivity has been demonstrated. In this paper, we have shown that the electromigrated nanoscale gaps that we use for single-molecule electronic transport experiments are extremely good for surface-enhanced Raman scattering (SERS). We show that one can make these structures in a scalable way, with a high yield, and they show all the hallmarks of few- or single-molecule sensitivity. We're very excited about where we might go with this....

Finally, this past weekend came the first announcement of results from Gravity Probe B, or as my friends and I liked to call it, "The Project that Ate Stanford". GPB is a satellite-based test of general relativity. My thesis advisor remembers Leonard Schiff coming to Cal Tech in 1964 and talking about how GPB was only a couple of years away. What does this have to do with condensed matter? Well, the folks building this thing produced a lot of good science trying to understand thin film niobium superconductors, including a technique using lead balloons (no, really) to produce volumes with magnetic fields smaller than anywhere else in the known universe. Anyway, the finally have announced some results - so far, they've found that GR does a good job (within 1%, anyway) at describing the bending of space near the earth. It would appear that the analysis of the main effect they're trying to see, "frame dragging" of spacetime by the rotation of the earth, is being hampered by annoying systematics in their experiment. Hopefully they'll get it worked out, though they suffer from a sociology of science problem: if they confirm GR, no one will be surprised; if they refute GR, no one may believe the result because the experiment is so complicated. Ahh well.

Friday, April 13, 2007

End of semester and a short break

Just a quick post.... We're approaching the end of the semester, and with several looming thesis defenses, a thesis prize committee commitment, an internal search for an administrative position, and various external deadlines, I will be cutting back on blogging for the next two or three weeks, with the possible exception of "This week in cond-mat" posts....

Wednesday, April 11, 2007

A paper and a wager

One paper on the arxiv that I need to read more closely: cond-mat/0701728, Sukhorukov et al., Conditional statistics of electron transport in interacting nanoscale conductors. This paper, a collaboration between Rochester, Geneva, and ETH Zurich, looks at the noise properties of transport through a quantum dot in the presence of a charge detector, in this case a quantum point contact (QPC). A QPC is a constriction in a 2d electron gas with a width modulated by gates. If the QPC is tweaked such that it's right on the boundary of pinching off a transverse electronic mode, its conductance can depend strongly on the local charge environment, which acts like an extra gate. By monitoring the conductance of the QPC, the authors can watch tunneling events in the capacitively coupled quantum dot. The presence of an electron on the dot reduces the conductance of the QPC. Like most of the very pretty work to come out of Enslinn's group at ETH, the quantum dot and QPC are defined through local AFM-based surface oxidation of a shallow 2d electron gas in a GaAs/AlGaAs heterostructure. What I don't understand about this paper is a statement in the introduction: "An important property of the QPC charge detector is its noninvasiveness: the system physically affects the detector, not visa-versa." Strictly speaking, this just can't be right. If the quantum dot is capacitively coupled to the QPC sufficiently to modulate the QPC current flow, there has to be back-action of the QPC on the dot charge. While that interaction may be small, it can't be nonexistent, as far as I can see.

An unrelated anecdote: some of you may remember this post, where I talked about Steorn, an Irish company that took out a full-page ad in the Economist magazine looking for scientists to act as a "jury" of some sort and evaluate their "free energy" machine. Well, it would appear that Steorn did manage to find some scientists willing to act as a "jury", whatever that means, and will release some sort of report of their findings this Friday. I actually did communicate with Steorn last fall; while they had some interest in talking to me, they did not come close to answering my questions about how their jury process was supposed to work (e.g., would scientists actually be able to play with the gadget in an off-site laboratory; did Steorn really mean that their machine produced energy, or were they trying to finesse conventional jargon by talking about "coefficients of performance greater than one" (which may mean nothing for a refrigerator, for example)). I'm willing to wager that they have not discovered a loophole in the first law of thermodynamics. It will be interesting to hear what they report, and whether any actual scientists would be willing to stand up and back Steorn's claims.

Update: Sean McCarthy of Steorn informs me that my comments above are inaccurate, and that, indeed, their evaluation will be based on tests specified by their "jury" in an independent laboratory. My apologies for any confusion. This was not clear to me last fall, and I haven't been following this in the interim. I should also point out that my interactions with them were entirely cordial and businesslike.

I reiterate, though, that I think there is zero chance that these folks have been able to circumvent conservation of energy. If they have, I will happily eat my words, as this would be the biggest science story of the century. Extraordinary claims require extraordinary evidence.

Saturday, April 07, 2007

Palate cleanser

Alright - after all of the recent seriousness, this blog needs a palate cleanser. How about a few fun links? For example, comic book artist and animator Neal Adams has some, umm, intriguing ideas about geophysics, such as the notion that the earth expanded greatly in diameter in the last 70 million years or so. Oh, and it's not just the earth. Mars, too.

I also recommend some Tom Lehrer music videos, if you haven't seen them before, like this, this, and this.

Friday, April 06, 2007

What I will and won't discuss, and comments

I'm annoyed that I have to make this post, but I guess that I'm going to have to actually set some sort of policy.

In the course of discussions related to the faculty job search process, comments have been made by posters regarding specific people and their job performance. Some of these comments have been about anonymous faculty or candidates; others have not. Some have been by anonymous posters; others have not.

Unfortunately (though understandably), Blogger does not allow me to go in and redact individual names from comments, as far as I know. I can either leave the comment alone or delete it altogether.

Consider this fair warning. If you post a comment that I think is inappropriate, I will delete it. In this case, I think anonymous comments that single out specific people (that is, where the identity of the subject of discussion is unambiguous, even if the full name isn't given) for criticism about their job performance and tenure chances are inappropriate. This policy isn't censorship - you're free to start your own blog and post whatever you want on it. I'm not going to provide a forum for those discussions.

UPDATE: Upon further reflection, I've deleted more comments from the previous thread, scrubbing out potential identifiers. I think I was able to preserve most of the insightful commentary while removing the tawdry gossip. I don't want to moderate posts, but I may have to go in that direction in the future. I wish blogger had an alternative to wiping out comments, but considering that the service is free, I shouldn't complain.

Wednesday, April 04, 2007

A primer on faculty searches, part III

Here is my long-delayed third post about faculty searches, a follow-up to Part I and Part II. One reason for the delay is that I was chairing a search. That took quite a bit of time, and I also didn't want to give an unfair advantage to any of our candidates who happened to read my blog.

In Part II I'd described the process in fairly complete detail. What I want to do here is give a few pointers to would-be candidates, and answer a couple of questions that people have emailed me in the interim.
  • Look over the department webpages before you visit. If the school gives you an advanced or draft copy of your schedule, actually look at the pages of the people you're going to meet. More than likely, most of the people you meet are on the search committee. You want to have some sense of what they do so that you can (a) ask decent questions as you meet them, and (b) pitch your own stuff at the right level. An astro person may not have any idea what "valley degeneracy" is.
  • Listen to what your point of contact tells you about who the audience is for your talk(s). No one wants a talk aimed at the wrong level. If you're supposed to give a general colloquium, that usually means that your audience is very broad and may contain undergrads, grad students, and faculty from different subfields. If you're supposed to give a seminar, that usually implies a more specialized audience. Remember, people need to know why they should care about what you're doing.
  • Rehearse your talk. Speak clearly. Do not speak super-fast, especially if you use technical terms or people's names.
  • Listen carefully to physics questions that you're asked, either in the talk or one-on-one. Repeat the question back at the person asking, rephrased slightly to confirm that you know what you're being asked. It's ok to say "I don't know" in response to a question, but don't use that dismissively. If you think of the answer later, make it a point to try to tell the questioner.
  • Don't use terms that you don't understand or can't explain. Don't assume that everyone in the audience has heard of the So-and-so Effect. Make sure that you know all the relevant numbers for your work. If you're a theorist and someone asks you how to measure the effect you're calculating, at least have a handwave idea. If you're an experimentalist and someone asks you about errors and uncertainties, make clear that you've thought about those issues.
  • When discussing budgets, etc., make sure that you know who actually is the point of contact for negotiations like that. In our case it's the department chairperson.
  • The rule on startup packages is generally "you might as well ask". Show some reasonable judgment, though. A junior person asking for $5M in startup is not reasonable. A junior person asking for 5000 sq.ft. of lab space is not reasonable.
  • Find out whether lab renovation costs are counted separately from your startup. That's the case at all the big schools, but some places can be funny about this.
  • Ask about the tenure process. Ask about tenure history in that department.
  • Ask about the department's long-term plan - where are things trending?
  • Find out what their schedule is. When do they think they will be wrapping up the search?
I've been asked about whether politics can enter decision making in these searches. Someone emailed me who had not been offered a position somewhere despite having multiple first-author papers, and the candidate that had been offered the position had many fewer and less postdoc experience. The short answer is, well-run searches make decisions based on the whole package (departmental needs; research quality; communications ability; personal interactions; etc.). So, it can be hard to say from the outside why a particular committee made particular decisions. Can there be poorly run searches? Can there be searches where one category or need trumps the others, and the candidates don't know about that, and isn't that unfair? Sure, and we don't do that because it's long-term stupid - that's not the way to hire for the future and build up a department. I'm not sure that the situation is worse in academia than in any other profession, though. There is no question that where someone comes from and what their "pedigree" is can have a big impact, as discussed here by the Ponderer, but the process is inherently complicated. Hopefully these postings have clarified things at least a little.

Monday, April 02, 2007

This week in cond-mat

Much as it pains me to admit this, I agree with Lubos Motl about something: Neither of us like the new numerical identifier system launched by Paul Ginsparg and company at the arxiv. Lubos nails both of my complaints. While the old system actually conveyed information (the subject area of the paper and how many papers in that category had been submitted that month), the new system manages to be both cryptic and uninformative. Frankly I don't care how many total papers have been submitted from all categories in the arxiv, and I'm not sure why anyone would. Somehow this reminds me of the apparent desire of the Powers that Be to switch NSF proposal submissions from FastLane, which works extremely well and is easy to use, to, which is completely arcane and annoying. Prof. Ginsparg, if you see this, please consider switching back to some incrementally changed form of the old system.

Meanwhile, here's one paper in the old numbering format, for old time's sake, that I thought looked interesting. Perhaps a theorist could take a look at this and tell me if it's as clever and neat as it seems to be.

cond-mat/0703768 - Ostlund, The strong coupling Kondo lattice model as a Fermi gas
The Kondo lattice is a model developed in an attempt to understand the heavy fermion compounds. In this model, there are itinerant conduction electrons, and a lattice of localized unpaired moments (f-shell electrons) representing the ion cores of the rare earth constituents of the heavy fermion material. Under the right conditions, the ground state of these materials is a Fermi liquid, meaning that there are distinct, gapless, electron-like (spin 1/2, charge -e) quasiparticles, but they have an effective mass hundreds of times higher than the free electron mass. The idea is that the conduction electrons have formed fully screened Kondo singlets with the rare earth f-electrons. The true ground state of the Kondo problem is a Fermi liquid, and in this limit the ground state of the Kondo lattice is also a Fermi liquid, though the antiferromagnetic screening of the ion cores leads to the high effective mass. Note that this is rather special - in semiconductors, the effective mass is a single-particle effect that comes from the lattice potential; in these systems, the effective mass is the result of many-body correlations. In this paper, the author explicitly writes down a canonical transformation (read: clever change of variables) that directly maps the Kondo lattice Hamiltonian into that of a weakly interacting Fermi gas. It looks clever to me, but I can't judge it in the context of other theoretical treatments of these strongly correlated systems.